相关论文: RAPO: Risk-Aware Preference Optimization for Gener…

Test-time scaling has proven effective in further enhancing the performance of pretrained Large Language Models (LLMs). However, mainstream post-training methods (i.e., reinforcement learning (RL) with chain-of-thought (CoT) reasoning)…

Large Reasoning Models (LRMs) introduce a new generation paradigm of explicitly reasoning before answering, leading to remarkable improvements in complex tasks. However, they pose great safety risks against harmful queries and adversarial…

Existing large language models (LLMs) face challenges of following complex instructions, especially when multiple constraints are present and organized in paralleling, chaining, and branching structures. One intuitive solution, namely…

Large Multimodal Reasoning Models (LMRMs) are moving into real applications, where they must be both useful and safe. Safety is especially challenging in multimodal settings: images and text can be combined to bypass guardrails, and single…

Large Reasoning Models (LRMs) improve task performance through extended inference-time reasoning. Although previous studies suggest that longer reasoning should lead to more robust safety behavior, we find evidence to the contrary:…

We reveal a critical yet underexplored flaw in Large Vision-Language Models (LVLMs): even when these models know the correct answer, they frequently arrive there through incorrect reasoning paths. The core issue is not a lack of knowledge,…

Medical question answering requires advanced reasoning that integrates domain knowledge with logical inference. However, existing large language models (LLMs) often generate reasoning chains that lack factual accuracy and clinical…

By integrating external knowledge, Retrieval-Augmented Generation (RAG) has become an effective strategy for mitigating the hallucination problems that large language models (LLMs) encounter when dealing with knowledge-intensive tasks.…

Aligning large language models with human preferences must balance two competing goals: responding helpfully to legitimate requests and reliably refusing harmful ones. Most preference-based safety alignment methods collapse safety into a…

Large Reasoning Models (LRMs) achieve remarkable success through explicit thinking steps, yet the thinking steps introduce a novel risk by potentially amplifying unsafe behaviors. Despite this vulnerability, conventional defense mechanisms…

Large Reasoning Models (LRMs) have demonstrated a latent capacity for complex reasoning by spontaneously exhibiting cognitive behaviors such as step-by-step reasoning, reflection, and backtracking, commonly referred to as "Aha Moments".…

The rapid advancement of multi-modal large reasoning models (MLRMs) -- enhanced versions of multimodal language models (MLLMs) equipped with reasoning capabilities -- has revolutionized diverse applications. However, their safety…

Multimodal Large Language Models (MLLMs) are susceptible to the implicit reasoning risk, wherein innocuous unimodal inputs synergistically assemble into risky multimodal data that produce harmful outputs. We attribute this vulnerability to…

While Reinforcement Learning (RL) shows promise in training tool-use Large Language Models (LLMs) using verifiable outcome rewards, existing methods largely overlook the potential of reasoning rewards based on chain-of-thought quality for…

Large Reasoning Models (LRMs) leverage transparent reasoning traces, known as Chain-of-Thoughts (CoTs), to break down complex problems into intermediate steps and derive final answers. However, these reasoning traces introduce unique safety…

Reasoning methods that adaptively allocate test-time compute have advanced LLM performance on easy to verify domains such as math and code. In this work, we study how to utilize this approach to train models that exhibit a degree of…

Recent advancements in Large Language Models (LLMs) have shifted from explicit Chain-of-Thought (CoT) reasoning to more efficient latent reasoning, where intermediate thoughts are represented as vectors rather than text. However, latent…

Adapting large language models (LLMs) for specific tasks usually involves fine-tuning through reinforcement learning with human feedback (RLHF) on preference data. While these data often come from diverse labelers' groups (e.g., different…

Large Reasoning Models (LRMs) generate structured chains of thought (CoTs) before producing final answers, making them especially vulnerable to knowledge leakage through intermediate reasoning steps. Yet, the memorization of sensitive…

Large Language Models (LLMs) have demonstrated remarkable capabilities but often face challenges with tasks requiring sophisticated reasoning. While Chain-of-Thought (CoT) prompting significantly enhances reasoning, it indiscriminately…